Background Quantifying joint deformity in people with rheumatoid arthritis (RA) and psoriatic arthritis (PsA), using high resolution peripheral quantitative computed tomography (HR-pQCT), remains problematic because it is difficult to estimate where the healthy joint surface would have been. Methods The second metacarpophalangeal of RA, PsA and healthy subjects were imaged with HR-pQCT. Using the bone surfaces of the healthy cohort as a reference, the method predicted the healthy surface of each individual diseased bone surface. Quantifiable outcomes were measured based on differences between the predicted healthy surface and the actual diseased surface. Sensitivity studies were conducted to measure precision, and the algorithm was validated against artificially created deformities with known geometries. Results Subjects with PsA and RA had significantly greater occurrences of erosion based surface outcomes than the healthy cohort. Sensitivity analyses revealed precision errors of up to 0.26 mm. Validating the algorithm showed an average accuracy error of 0.12 mm (4%) for detecting erosions and 0.27 mm (20%) for detecting periosteal bone growths. Conclusions The new method allows for visualization and quantification of surface changes within the affected joint by identifying areas of erosion and periosteal bone formation. Surface based outcomes are a novel way to interpret and further quantify articular bone changes affected by PsA and RA.